基于改进分形技术的红外图像噪音抑制研究

Infrared image noise repressionresearch based on improved fractal

红外图像杂噪音抑制是图像识别的重要技术基础。阐述如何对传统分形技术进行改进并将多重分形理论引入红外图像的噪音抑制,提出一种基于奇异性分析的红外图像消噪方法。在图像经过多重分形分析后熏通过对多重分形谱相关的Holder指数进行特征化处理来实现图像消噪。这种方法可以在尽量保留信号奇异性穴如边缘熏纹理等雪相关信息情况下熏使得图像平滑熏噪声消除,是红外图像消噪的一种新方法。

Infrared image noise repression is an important technique foundation of image interpretation. This paper improves the traditional fractal technique, introduces multifractal theory into infrared image noise repression, brings forward an infrared image noise repression method based on odd analyse. After multifractal analyse, noise repression can be realized by characterizing processing of Holder index related to multifractal. This method can repress noise in reserving the related information, such as edge and vein etc. It is a new method of infrared image noise repression.